Cannabis yield, potency, and leaf photosynthesis respond

Cannabis yield, potency, and leaf photosynthesis respond differently to 1

increasing light levels in an indoor environment 2

Victoria Rodriguez Morrison†, David Llewellyn†, and Youbin Zheng* 3

School of environmental Sciences, University of Guelph, Guelph, Ontario, Canada, N1G 2W1 4

* Correspondence: 5

Youbin Zheng 6

[email protected] 7

Keywords: Cannabis sativa, PPFD, light intensity, light response curve, indoor, sole source, 8

cannabinoid, terpene 9

Abstract 10

Since the recent legalization of medical and recreational use of cannabis (Cannabis sativa L.) in 11

many regions worldwide, there has been high demand for research to improve yield and quality. With 12

the paucity of scientific literature on the topic, this study investigated the relationships between light 13

intensity (LI) and photosynthesis, inflorescence yield, and inflorescence quality of cannabis grown in 14

an indoor environment. After growing vegetatively for 2 weeks under a canopy-level photosynthetic 15

photon flux density (PPFD) of ≈ 425 μmol·m-2·s-1 and an 18-h light/6-h dark photoperiod, plants 16

were grown for 12 weeks in a 12-h light/12-h dark ‘flowering’ photoperiod under canopy-level 17

PPFDs ranging from 120 to 1800 μmol·m-2·s-1 provided by light emitting diodes. Leaf light response 18

curves varied both with localized (i.e., leaf-level) PPFD and temporally, throughout the flowering 19

cycle. Therefore, it was concluded that the leaf light response is not a reliable predictor of whole-20

plant responses to LI, particularly crop yield. This may be especially evident given that dry 21

inflorescence yield increased linearly with increasing canopy-level PPFD up to 1800 μmol·m-2·s-1, 22

while leaf-level photosynthesis saturated well below 1800 μmol·m-2·s-1. The density of the apical 23

inflorescence and harvest index also increased linearly with increasing LI, resulting in higher-quality 24

marketable tissues and less superfluous tissue to dispose of. There were no LI treatment effects on 25

cannabinoid potency, while there were minor LI treatment effects on terpene potency. Commercial 26

cannabis growers can use these light response models to determine the optimum LI for their 27

production environment to achieve the best economic return; balancing input costs with the 28

commercial value of their cannabis products. 29

1 INTRODUCTION 30

Drug-type Cannabis sativa L. (hereafter, cannabis) is often produced indoors to allow complete 31

control of environmental conditions, which is important for producing consistent medicinal plants 32

and products (United Nations Office on Drugs and Crime, 2019; Zheng, 2020). Total reliance on 33

electrical lighting for plant production gives growers the capability to manipulate crop morphology, 34

yield, and quality using light. However, lighting-related costs comprise ≈ 60% of total energy used 35

for indoor cannabis production (Evergreen Economics, 2016; Mills, 2012); making crop lighting one 36

of the most substantial input costs for growing cannabis indoors. With recent nationwide legalization 37

in Canada (among many other regions worldwide), energy demand for indoor cannabis production is 38

expected to increase rapidly as the industry intensifies production to address rising demand (Sen and 39

Wyonch, 2018). 40

Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 8 January 2021 doi:10.20944/preprints202101.0163.v1

© 2021 by the author(s). Distributed under a Creative Commons CC BY license.

https://doi.org/10.20944/preprints202101.0163.v1

http://creativecommons.org/licenses/by/4.0/

2

There are many factors that govern the cost of producing photosynthetically active radiation (PAR) 41

for indoor cannabis production. These factors include: the capital and maintenance costs of lighting 42

fixtures and related infrastructure, efficiency of converting electricity into PAR (usually referred to as 43

PAR efficacy; in units of µmol(PAR)·J-1), management of excess heat and humidity, and uniformity of 44

PAR distribution within the plant canopy. The most common lighting technologies used for indoor 45

cannabis production are high intensity discharge (e.g., high pressure sodium) and light emitting 46

diodes (LED). These technologies have widely varying spectrum, distribution, PAR efficacy, and 47

capital costs. However, regardless of the lighting technology used, the dominant factor that regulates 48

the cost of crop lighting is the target canopy-level light intensity (LI). 49

One common precept in controlled-environment agriculture production is that crop yield responds 50

proportionally to increasing LI; i.e. the so-called “1% rule” whereby 1% more PAR equals 1% 51

greater yield (Marcelis et al., 2006). On a per-leaf basis, this principle is clearly limited to lower light 52

intensities, since light use efficiency (i.e., maximum quantum yield; QY, μmol(CO2)·μmol-1(PAR)) of 53

all photosynthetic tissues begins to decline at LI well below their light saturation points (LSP; i.e., 54

the LI at peak photosynthetic rate) (Posada et al., 2012). However, in indoor-grown cannabis, it is 55

conceivable that whole-plant photosynthesis will be maximized when LI at the upper canopy leaves 56

are near their LSP. This is partly attributable to the inter-canopy attenuation of PAR from self-57

shading; allowing lower-canopy foliage to function within the range of LIs where their respective 58

LUE are optimized (Terashima and Hikosaka, 1995). This may be especially relevant to indoor 59

production, where relatively small changes in distance from the light source can impart substantial 60

differences in foliar LI (Niinemets and Keenan, 2012). Further, distinguished from many other 61

indoor-grown crops, cannabis foliage appears to tolerate very high LI, even when exposed to 62

photosynthetic photon flux densities (PPFD) that are much higher than what they have been 63

acclimated to (Chandra et al., 2015). 64

There is a paucity of peer-reviewed studies that have related LI to cannabis potency and yield (e.g., 65

mass of dry, mature inflorescence per unit area and time). Perhaps the most referenced studies 66

reported aspects of single-leaf photosynthesis of several cultivars and under various PPFD, CO2 67

concentration, and temperature regimes (Chandra et al., 2011; 2015; Lydon et al., 1987). These 68

works have demonstrated that cannabis leaves have very high photosynthetic capacity. However, 69

they have limited use in modeling whole canopy photosynthesis or predicting yield because single-70

leaf photosynthesis is highly variable; depending on many factors during plant growth such as: leaf 71

age, their localized growing environments (e.g., temperature, CO2, and lighting history), and 72

ontogenetic stage (Bauerle et al., 2020; Carvalho et al., 2015; Murchie et al., 2002; Zheng et al., 73

2006). While lighting vendors have long relied on cannabis leaf photosynthesis studies to sell more 74

light fixtures to cannabis growers, their models are only tangentially related to whole-canopy 75

photosynthesis, growth, and (ultimately) yield (Kirschbaum, 2011). 76

Some forensic studies have utilized various methods to develop models to estimate crop yield from 77

illicit indoor cannabis production (Backer et al., 2019; Potter and Duncombe, 2012; Toonen et al., 78

2006; Vanhove et al., 2011). These models used an array of input parameters (e.g., planting density, 79

growing area, crop nutrition factors, etc.) but, they relied on “installed wattage” (i.e., W·m-2) as a 80

proxy for LI. It is notable that reporting yield as g·W-1 (i.e., g·m-2 / W·m-2) overlooks the 81

instantaneous time factor inherent in power units (i.e., W = J·s-1). A more appropriate yield metric 82

would also account for the length of the total lighting time throughout the production period (i.e., h·d-83 1 × d), thus factoring out the time units resulting in yield per unit energy input (e.g., g·kWh-1). 84

Further, area-integrated power does not directly correlate to the canopy-level light environment due 85



3

to myriad unknowns, such as hang height, light distribution, and fixture efficacy. It is therefore 86

impossible to accurately ascertain canopy-level LI in these models. Eaves et al. (2020) reported linear 87

relationships between canopy-level LI (up to 1500 µmol·m-2·s-1) and yield; however, they had only 88

one LI treatment above 1000 µmol·m-2·s-1. Further, they reported substantial inter-repetition 89

variability in their yield models, which indicates that factors other than LI may have limited crop 90

productivity in some circumstances. While methodological deficiencies in these studies may limit the 91

confident quantitative extrapolation of their results to production environments, it is striking that 92

none of these studies reported evidence of saturation of inflorescence yield at very high LI. 93

These studies all demonstrate the exceptionally high capacity that cannabis has for converting PAR 94

into biomass. However, there are also clear knowledge gaps in cannabis’ photosynthesis and yield 95

responses to increasing LI. Further, cannabis products are very high-value commodities relative to 96

other crops grown in indoor environments. This means that producers may be willing to accept 97

substantially higher lighting-related input costs in order to promote higher yields in limited growing 98

areas. However, maximizing yield regardless of cost is not a feasible business model for most 99

cannabis producers; rather there is a trade-off between input costs and crop productivity by selecting 100

the optimum canopy-level LI (among other inputs) that will maximize net profits. Further 101

complicating matters, producers must balance fixed costs which do not vary with crop productivity 102

(such as property tax, lease rates, building security, and maintenance, etc.) and variable costs (such as 103

the aforementioned lighting-related costs among other crop inputs) which can have dramatic impacts 104

on crop productivity and yield (Vanhove et al., 2014). Since indoor crop lighting is a compromise 105

between input costs and crop productivity, it is critical for growers to select the optimum light 106

intensity (LI) for their respective production environment and business models. 107

The objectives of this study were to establish the relationships between canopy-level LI, leaf-level 108

photosynthesis, and yield and quality of drug-type cannabis. We investigated how plant growth stage 109

and localized foliar PPFD (LPPFD; i.e., instantaneous PPFD at leaf-level) affected photosynthetic 110

parameters and leaf morphology, and how growing cannabis at average canopy-level PPFDs 111

(APPFD; i.e., lighting history) ranging from 120 to 1800 µmol·m-2·s-1 affected plant morphology, 112

yield, and quality of mature marketable inflorescence. The results of this study will assist the indoor 113

cannabis industry to determine how much PAR cannabis growers should be providing to the crop 114

canopy in order to maximize profits while minimizing energy use within their specific production 115

scenarios. 116

2 MATERIAL AND METHODS 117

The trial area consisted of 2 adjacent deep-water culture basins (CB) located in an indoor cannabis 118

production facility in Southern Ontario, Canada. Each CB (14.6 x 2.4 m) consisted of 24 parallel 119

polystyrene rafts (0.6 x 2.4 m), each containing holes for 16 plant pots, oriented in 2 rows with 30-cm 120

spacing both within- and between-rows. This spacing provided for 384 plants to be evenly spaced 121

within each CB, at a density of 0.09 m2/plant. 122

Above each CB were 3 racks of LED fixtures (Pro-650; Lumigrow, Emeryville, CA, USA), with 123

each rack consisting 2 rows of 4 fixtures each; arranged such that all 24 fixtures were uniformly-124

spaced (1.2 m apart, on-center) relative to each other and centered over the footprint of the CB. Each 125

rack of fixtures was height-adjustable via a system of pulleys and cables, such that the hang-height of 126

the 8 fixtures in each rack could be adjusted in unison. Each fixture contained dimmable spectrum 127

channels for blue (B, peak 455 nm), white (broad-spectrum 5000K) and red (R, peak 660 nm) which 128



4

could be individually controlled, wirelessly, through Lumigrow’s SmartPAR software. The photon 129

flux ratio of B (400-500 nm), green (G, 500-600 nm), and R (600-700 nm) was B18:G5:R77. 130

Relative spectral photon flux distribution (Figure 1) was measured using a radiometrically calibrated 131

spectrometer (UV-VIS Flame-S-XR; Ocean Optics, Dunedin, FL, USA) coupled to a CC3 cosine-132

corrector attached to a 1.9 m x 400 µm UV-Vis optical fibre. 133

134

Figure 1. Relative spectral photon flux distribution of Pro-650 (Lumigrow) light-emitting diode 135

(LED) fixtures. 136

2.1 Experimental Design 137

The experiment was conducted using a gradient design, whereby plants grown in a common 138

environment were exposed to a broad range of canopy-level PPFDs with a high level of spatial 139

variability across the CB. Individual plants were assigned APPFD levels based on rigorous spatial 140

and temporal evaluations of LI (explained below). Gradient designs can outperform traditional 141

“treatment x replication” experimental designs when evaluating plants’ responses to a continuous 142

variable such as LI (Kreyling et al., 2018). While they are arduous to setup and monitor, gradient 143

designs have been successfully used to establish LI effects within other controlled-environment 144

production scenarios (Bredmose, 1993, 1994; Jones-Baumgardt et al., 2019). 145

At its outset, the experiment was arranged as a randomized complete block design (RCBD) with 6 146

blocks of 8 PPFD target levels: 200, 400, 600, 800, 1000, 1200, 1400, and 1600 μmol·m-2·s-1, to 147

facilitate setup. Each block consisted of a single rack of LED fixtures, with the PPFD target levels 148

randomly assigned to individual fixtures (i.e., plots) within each rack. The two plants located most 149

directly below each fixture were assessed experimentally. PPFD was measured at the apex of each 150



5

plant using a portable spectroradiometer (LI-180; LI-COR Biosciences, Lincoln, NE, USA). The 151

initial hang height of each rack was determined by the maximum height whereby approximately 1600 152

μmol·m-2·s-1 could be achieved at the apical meristem of the tallest plant in the highest LI plot. The 153

other treatment levels were subsequently achieved through dimming; targeting the prescribed PPFD 154

at the apical meristem of the tallest plant in each plot while maintaining a uniform photon flux ratio 155

of B18:G5:R77 in the entire CB. Plant height and apical meristematic PPFD were measured twice 156

weekly until vegetative growth ceased (five weeks after the start of the 12-h photoperiod), and 157

weekly thereafter until harvest. The prescribed intensity levels in each block were reset each time 158

plant height was measured, first by raising the rack of fixtures to achieve the target PPFD at the 159

apical meristem of the tallest plant in the 1600 μmol·m-2·s-1 plot and then adjusting the intensity 160

settings of the remaining plots accordingly. The trial ran from the beginning of the flowering stage 161

(i.e., when the 12-h flowering photoperiod was initiated) until harvest, for a total of 81 days (nearly 162

12 weeks). 163

While the underlying experimental arrangement was based on a RCBD organization, all analyses 164

were performed as regressions with LI as the continuous, independent variable. 165

2.2 PPFD Levels 166

While the prescribed target PPFD levels were maintained at the apical meristem at the tallest plant 167

within each plot on regular intervals, these values were not accurate proxies for the actual PPFD 168

intensity dynamics experienced by each plant throughout the trial due to variability in individual 169

plant height (on intra- and inter-plot bases), growth rates, and the lengths of the time periods between 170

PPFD measurements. To account for these temporal dynamics in apical meristematic PPFD, total 171

light integrals (TLIs, mol·m-2) were calculated for each plant over the total production time and then 172

back-calculated to APPFD or daily light integral (DLI, mol·m-2·d-1). The TLIs were based on the 173

product of the average PPFD level measured at the start and end of each measurement interval and 174

the length of time the lights were on during each measurement interval. These interim light integrals 175

were then aggregated to form a TLI for each plant and divided by the total production time in 176

seconds (i.e., the product of the daily photoperiod and the number of days). The resulting APPFD 177

levels were then used as the independent variable (i.e., X-axis) in regressions of LI vs. various 178

growth, yield and quality parameters. TLI can also be used in yield evaluations whereby the 179

relationship between yield and TLI becomes a direct measure of production efficacy on a quantum 180

basis (e.g., g·mol-1). This relationship can be converted to an energy-basis (g·kWh-1), if the fixture 181

efficacy (μmol·J-1) and spatial distribution efficiency (i.e., proportion of photon output from fixtures 182

that reach the target growing area) are known. 183

2.3 Plant Culture 184

Cuttings were taken from mother plants of the ‘Stillwater’ cultivar on 1 Aug. and 15 Aug. 2019 and 185

rooted in stone wool cubes under 100 μmol·m-2·s-1 of fluorescent light for 14 d and then transplanted 186

into a peat-based medium in 1-gallon plastic pots and grown under ≈ 425 μmol·m-2·s-1 of LED light, 187

comprised of a mixture of Pro-325 (Lumigrow) and generic phosphor-converted white LEDs 188

(unbranded) for an additional 14 d. The apical meristems were removed (i.e., “topped”) from the first 189

batch of clones, 10 d after transplant, and the second batch were not topped. Propagation and 190

vegetative growth phases both had 18-h photoperiods. The first CB (CB1) was populated from the 191

first batch of clones on 29 Aug. 2019 and the second CB (CB2) was populated from the second batch 192

of clones on 12 Sept. 2019. In each case, 48 uniform and representative plants were selected from the 193

larger populations of clones and placed in the plots to be evaluated experimentally. In CB1, the 194



6

experimental plants initially had either 9 or 10 nodes and ranged in height (from growing medium 195

surface to shoot apex) from 34 to 48 cm. In CB2 the experimental plants initially had either 12 or 13 196

nodes and ranged in height from 41 to 65 cm. Once the plants were moved to the CBs, the daily 197

photoperiod switched to 12 h, from 06:30 HR to 18:30 HR. 198

Plant husbandry followed the cultivator’s standard operating procedures except for the differences in 199

canopy-level PPFD. Canopy-level air temperature, relative humidity (RH), and carbon dioxide (CO2) 200

concentration were monitored on 600-s intervals throughout the trial with a logger (Green Eye model 201

7788; AZ Instrument Corporation, Taiwan). The air temperature, RH, and CO2 concentrations were 202

(mean ± SD) 25.3 ± 0.4 °C, 60.5 ± 4.8%, and 437 ± 39 ppm during the day (i.e., lights on) and 25.2 ± 203

0.3 °C, 53.1 ± 3.3%, and 479 ± 42 ppm during the night. A common nutrient solution is circulated 204

through both CBs. The nutrient concentrations in the aquaponic solution were sampled weekly and 205

analyzed at an independent laboratory (A&L Canada; London, ON, Canada). The nutrient element 206

concentrations (mg·L-1) in the aquaponic system were (mean ± SD, n = 11): 170 ± 22 Ca, 86 ± 8.2 S, 207

75 ± 15 N, 57 ± 5 Mg, 32 ± 4 P, 23 ± 8 K, 250 ± 32 Cl, 0.27 ± 0.1 Fe, 0.18 ± 0.07 Zn, 0.050 ± 0.02 208

Mn, 0.031 ± 0.006 B, and 0.028 ± 0.004 Cu. Mo was reported as below detection limit (i.e., < 0.02 209

mg·L-1) throughout the trial. The concentrations (mg·L-1) of non-essential nutrient elements were 170 210

± 18 Na and 6.7 ± 0.7 Si. The aquaponic solution was aerated with an oxygen concentrator and the 211

pH and EC were 6.75 ± 0.2 and 1.77 ± 0.15 mS·cm-1, respectively. 212

2.4 Leaf Photosynthesis 213

Quantifications of leaf-level gas exchange of leaflets on the youngest, fully-expanded fan leaves were 214

performed on 64 plants (32 plants per CB) each, in weeks 1, 5, and 9 after the initiation of the 12-h 215

photoperiod using a portable photosynthesis machine (LI-6400XT; LI-COR Biosciences), equipped 216

with the B and R LED light source (6400-02B, LI-COR Biosciences). The Light Curve Auto-217

Response subroutine was used to measure net carbon exchange rate (NCER; μmol(CO2)·m-2·s-1) at 218

PPFD levels of: 2000, 1600, 1400, 1200, 1000, 800, 600, 400, 200, 150, 100, 75, 50, 25, and 0 219

µmol·m-2·s-1. Leaflets were exposed to 2000 µmol·m-2·s-1 for 180 s prior to starting each light 220

response curve (LRC) and then progressed sequentially from highest to lowest PPFD to ensure 221

stomatal opening was not a limitation of photosynthesis (Singsaas et al., 2001). The leaf chamber 222

setpoints were 26.7°C (block temperature), 400 ppm CO2, and 500 µmol·s-1 airflow. The localized 223

PPFD (LPPFD) at each leaflet was measured immediately prior to the LRC measurement using the 224

LI-180. The light-saturated net CO2 exchange rate (Asat; μmol(CO2)·m-2·s-1), localized NCER 225

(LNCER; i.e., the NCER at LPPFD), maximum quantum yield (QY; μmol(CO2)·μmol-1(PAR), and 226

light saturation point (LSP; μmol(PAR)·m-2·s-1) were determined for each measured leaflet using 227

Prism (Version 6.01; GraphPad Software, San Diego, CA, USA) with the asymptotic LRC model: y 228

= a + b·e(c·x) (Delgado et al., 1993) where y, x, a, and e represent NCER, PPFD, Asat, and Euler’s 229

number, respectively. The LNCER of each leaflet was calculated by substituting the measured 230

LPPFD into its respective LRC model. The QY was calculated as the slope of the linear portion of the 231

LRC (i.e., at PPFD ≤ 200 μmol·m-2·s-1). The LSP is defined as the PPFD level where increasing LI 232

no longer invokes a significant increase in NCER. The LSP for each LRC was determined using the 233

methods described by Lobo et al. (2013) by evaluating the change in NCER (ΔNCER) over 50 234

μmol(PAR)·m-2·s-1 increments, continuously along the LRC, until the ΔNCER reached a threshold 235

value, which was determined from the prescribed measurement conditions and performance 236

specifications of the LI-6400XT. Briefly, the minimum significant difference in CO2 concentration 237

between sample and reference measurements is 0.4 ppm (LI-COR Biosciences, 2012). Therefore, 238



7

given the setup parameters of the leaf chamber, a ΔNCER of ≤ 0.33 μmol(CO2)·m-2·s-1 over a 50 239

µmol(PAR)·m-2·s-1 increment indicated the LSP. 240

The ratio of variable to maximum fluorescence (Fv/Fm) emitted from photosystem II (PSII) in dark-241

acclimated leaves exposed to a light-saturating pulse is an indicator of maximum quantum yield of 242

PSII photochemistry (Murchie and Lawson, 2013). Immediately after each LRC, the leaflet was dark 243

acclimated for ≈ 900 s and then Fv/Fm was measured with a fluorometer (FluorPen FP 100; Drasov, 244

Czech Republic). Chlorophyll content index (CCI) was measured on three fan leaflets from leaves at 245

the bottom and top of each plant in weeks 1, 5, and 9 using a chlorophyll meter (CCM-200; Opti-246

Sciences, Hudson, NH, USA). The CCI measurements from upper and lower tissues, respectively, 247

were averaged on a per-plant basis for each measurement period. 248

2.5 Leaf Morphology 249

On day 35, one leaf from each plant was removed from node 13 (counting upwards from the lowest 250

node) in CB1 and node 15 from CB2, ensuring that the excised leaves developed under their 251

respective LPPFD. A digital image of each leaf was taken using a scanner (CanoScan LiDE 25; 252

Canon Canada Inc., Brampton, ON, Canada) at 600 dpi resolution and then the leaves were oven-253

dried (Isotemp Oven Model 655G; Fisher Scientific, East Lyme, CT, USA), singly, to constant 254

weight at 65°C. The images were processed using ImageJ 1.42 software (National Institute of Health; 255

https://imagej.nih.gov/ij/download.html) to determine leaf area (LA). The dry weights (DW) of 256

scanned leaves were measured using an analytical balance (MS304TS/A00; Mettler-Toledo, 257

Columbus, OH, USA). Specific leaf weight (SLW; g·m-2) was determined using the following 258

formula: DW / LA. 259

2.6 Yield and Quality 260

After 81 d, the stems of each plant was cut at substrate level and the aboveground biomass of each 261

plant was separated into three parts: apical inflorescence, remaining inflorescence, and stems and 262

leaves (i.e., non-marketable biomass), and weighed using a digital scale (Scout SPX2201; OHAUS 263

Corporation, Parsippany, NJ, USA). Since the plants from CB2 had the apical meristem removed, the 264

inflorescence from the tallest side branch was considered the apical inflorescence. The length (L) and 265

circumference (C; measured at the midpoint) of each apical inflorescence were also measured. 266

Assuming a cylindrical shape, the density of the apical inflorescence (AID, g·cm-3) was calculated 267

using the formula: AID = fresh weight/{π·[C/(2·π )2]·L}. The apical inflorescences from 22 268

representative plants from CB1 were air dried at 15 °C and 40% RH for 10 d until they reached 269

marketable weight (i.e., average moisture content of ≈ 11%), determined using a moisture content 270

analyzer (HC-103 Halogen Moisture Analyzer; Mettler-Toledo, Columbus, OH, USA). This ensured 271

that the apical inflorescence tissues selected for analysis of secondary metabolites followed the 272

cultivator’s typical post-harvest treatment. The apical inflorescences from CB1 were homogenized on 273

a per-plant basis and ≈ 2-g sub-samples from each plant was processed by an independent laboratory 274

(RPC Science & Engineering; Fredericton, NB, Canada) for potency (mg·g-1(DW)) of 11 cannabinoids 275

and 22 terpenes using ultra-high-performance liquid chromatography and mass spectrometry. Total 276

equivalent Δ-9-tetrahydrocannabinol (∆9-THC), cannabidiol (CBD), and cannabigerol potencies were 277

determined by assuming complete carboxylation of the acid-forms of the respective cannabinoids, 278

whose concentrations were adjusted by factoring out the acid-moiety from the molecular weight of 279

each compound [e.g., total ∆9-THC = (∆9-THCA x 0.877) + ∆9-THC]. The separated aboveground 280

tissues from 16 representative plants in each CB were oven-dried (Isotemp Oven Model 655G) to 281



8

constant weight at 65°C to determine LI treatment effects on moisture content, which were then used 282

to determine DW of all harvested materials. The harvest index (HI) was calculated as the ratio of 283

total inflorescence DW (hereafter, yield) to the total aboveground DW, on a per-plant basis. 284

2.7 Data Processing and Analysis 285

On per-CB and per-week bases, each model from the leaf photosynthesis measurements (i.e., Asat, 286

LSP, LNCER, and QY) were subjected to non-linear regression using the PROC NLMIXED 287

procedure (SAS Studio Release 3.8; SAS Institute Inc., Cary, NC), with the LPPFD of each measured 288

leaf as the independent variable, to determine the best-fit models after outliers were removed. In 289

each case, best-fit models were selected based on the lowest value for the Akaike information 290

criterion (AICc). If there were no LI treatment effects on a given parameter, then means (± SD) were 291

calculated. Best-fit models for Fv/Fm and CCI were similarly determined, using LPPFD and APPFD 292

(from the start of the trial up to the time of measurement), respectively, as the independent variable. 293

On a per-week basis, Asat, LSP, LNCER, QY, Fv/Fm, and CCI data from CB1 and CB2 were pooled if 294

the 95% confidence intervals (95% CI) of each element of the respective best-fit models for the two 295

CBs overlapped, and best-fit models for pooled datasets were then recalculated. The PROC 296

GLIMMIX Tukey-Kramer test was used (P ≤ 0.05) on the resulting models (including means) to 297

determine if there were differences between the measurement periods (i.e., weeks). If there were any 298

measurement period effects on any element in the models, then weekly models for the respective 299

parameters were reported. 300

Computed parameters from single-time measurements (SLW, AID, yield, and HI) were grouped per 301

CB, using the APPFD (at the time of measurement) to define each datapoint within each CB and 302

PROC NLMIXED was used to evaluate the best fit model for each parameter using the AICc. 303

Parameter means were computed (on per-CB bases) when there were no LI treatment effects. If there 304

were LI treatment effects on a given parameter, datasets from CB1 and CB2 were pooled if the 95% 305

confidence intervals (95% CI) of each element of the respective best-fit models for the two CBs 306

overlapped and best-fit models for pooled datasets were then recalculated. For parameters with no LI 307

treatment effects, differences between CBs were evaluated using the 95% CI’s of their respective 308

means. For a given parameter, if the 95% CIs the parameter means for the 2 CBs overlapped, then the 309

data was pooled and new parameter means were calculated and presented. Cannabinoids and terpenes 310

from CB1 were modeled, with APPFD as the independent variable, using PROC NLMIXED to 311

evaluate the best-fit model for each parameter using the AICc. Best-fit models or parameter means 312

were reported. 313

3 RESULTS 314

No CB effects were found in any leaf photosynthesis, leaf morphology, and post-harvest parameters; 315

therefore, CB1 and CB2 data were pooled for the development of all models except secondary 316

metabolites, which were only measured in CB1. In contrast, many of the parameters that were 317

repeated over time (i.e., in weeks 1, 5, and 9) showed differences between weeks; whereby the 318

different weeks were modeled separately. Note also that the week-over-week ranges of LPPFD varied 319

as the plants progressed through their ontogeny, since self-shading from upper tissues resulted in 320

decreases in maximum LPPFD of leaves selected for photosynthesis measurements. Nevertheless, a 321



9

consistent range of APPFDs range was maintained throughout the trial. 322

323

Figure 2. Typical light response curves [net CO2 exchange rate (NCER) response to light intensity] 324

of the youngest fully-expanded fan leaves of Cannabis sativa L. ‘Stillwater’ grown under either low 325

or high localized photosynthetic photon flux densities (LPPFD). The low and high LPPFD were 91 326

and 1238 μmol·m-2·s-1, respectively. Measurements were made during week 5 after the initiation of 327

the 12-h photoperiod. 328



10

329

Figure 3. The light-saturated net CO2 exchange rate (Asat) (A), the light saturation point (LSP) (B), 330

the localized net CO2 exchange rate (LNCER) (C), and the Fv/Fm (D) of the youngest fully-expanded 331

fan leaves of Cannabis sativa L. ‘Stillwater’ at the localized photosynthetic photon flux densities 332

(LPPFD) that the respective leaves were growing under when the measurements were made, during 333

weeks 1, 5, and 9 after initiation of the 12-h photoperiod. Each datum is a single plant. Regression 334

lines are presented when P ≤ 0.05. 335

336

3.1 Leaf Photosynthesis 337

Leaf light response curves constructed under different LI and at different growth stages (week 1, 5, 338

and 9) generally demonstrated the trends that the Asat and LSP were higher for plants grown under 339

high vs. low LPPFD (Figures 2, 3A-B), especially after the plants had acclimated to their new 340

lighting environments (i.e., weeks 5 and 9). There were no LPPFD effects on Asat in week 1, with a 341

mean (± SE, n = 52) of 23.9 ± 0.90 μmol(CO2)·m-2·s-1 (Figure 3A). The Asat in weeks 5 and 9 (Figure 342

3A) and LSP in weeks 1, 5, and 9 (Figure 3B) increased linearly with increasing LPPFD. At low 343

LPPFD, the highest LSP was in week 1. The slopes of the Asat and LSP models were similar in weeks 344

5 and 9, but the Y-intercepts for both parameters were approximately twice as high in week 5 vs. 345

week 9. LNCER increased linearly with increasing LPPFD in weeks 1, 5, and 9 (Figure 3C) with the 346

steepest and shallowest slopes coming in weeks 5 and 1, respectively. The LNCER model in week 9 347

had a substantially lower Y-intercept than the other two weeks. As evidenced by the projected 348



11

intersection of the Asat and LNCER models in week 5 (i.e., at LPPFD of 1532 μmol·m-2·s-1), the 349

maximum LPPFD in week 5 (i.e., 1370 μmol·m-2·s-1) was nearly sufficient to saturate the 350

photosynthetic apparatus at the top of the canopy. There were no LPPFD effects on QY, but the mean 351

QY in weeks 1 and 5 were higher than week 9. The mean (± SE) QY were 0.066 ± 0.0013 (n = 54), 352

0.068 ± 0.0005 (n = 60), and 0.058 ± 0.0008 (n = 63) μmol(CO2)· μmol-1(PAR) in weeks 1, 5, and 9 353

respectively. The Fv/Fm decreased linearly with increasing LPPFD in all three measurement periods 354

(Figure 3D). The Fv/Fm model from week 5 had the largest Y-intercept (0.832) but also the steepest 355

slope. 356

357

Figure 4. The specific leaf weight (SLW; on a dry weight basis) of young, fully-expanded Cannabis 358

sativa L. ‘Stillwater’ leaves in response to the average photosynthetic photon flux density (APPFD), 359

measured on day 35 after initiation of the 12-h photoperiod. Each datum represents one fan leaf from 360

a single plant. 361



12

362

Figure 5. Sketches of Cannabis sativa L. ‘Stillwater’ plants grown under low (A) and high (B) 363

photosynthetic photon flux density (APPFD), 9 weeks after initiation of 12-h photoperiod (illustrated 364

by Victoria Rodriguez Morrison). 365

366

3.2 Chlorophyll Content Index and Plant Morphology 367

There were no LI treatment effects on CCI either at the top or bottom of the canopy, however within 368

in each week, the upper canopy CCI were higher than the lower canopy. Additionally, the CCI in the 369

upper and lower canopy was higher in week 1 vs. weeks 5 and 9. The CCI (means ± SE, n = 91) were 370

67.1 ± 0.80, 55.8 ± 2.2, and 52.0 ± 2.1 in the upper canopy and 46.3 ± 1.1, 31.1 ± 0.86, and 31.5 ± 371

1.1 in the lower canopy, in weeks 1, 5, and 9 respectively. The SLW increased linearly from 35.4 to 372

58.1 g·m-2 as APPFD (calculated based on the respective plants’ accumulated PAR exposures up to 373

day 35 of the flowering stage) increased from 130 to 1990 μmol·m-2·s-1 (Figure 4). Plants grown 374

under low vs. high APPFD were generally shorter and wider, with thinner stems, larger leaves, and 375

fewer, smaller inflorescences (Figure 5). 376



13

377

Figure 6. The relationship between average apical photosynthetic photon flux density (APPFD) 378

applied during the flowering stage (81 days) and inflorescence dry weight (A), harvest index (HI; 379

total inflorescence dry weight / total aboveground dry weight) (B), and apical inflorescence density 380

(AID; based on fresh weight) (C) of Cannabis sativa L. ‘Stillwater’. Each datum is a single plant. 381

Table 1. Cannabinoid potency in apical inflorescences of Cannabis sativa L. ‘Stillwater’. 382

383



14

Cannabinoid Potency (mg·g-1 of

inflorescence dry weight)

Δ-9-tetrahydrocannabinol (Δ9-THC) UDLz

Δ-9-tetrahydrocannabinolic Acid (Δ9-THCA) 12.9y ± 0.03

Total equivalent Δ9-tetrahydrocannabinol (TΔ9-THC) 11.3 ± 0.02

Cannabidiol (CBD) 5.53 ± 0.01

Cannabidiolic acid (CBDA) 214 ± 0.4

Total equivalent cannabidiol (TCBD) 193 ± 0.4

Cannabigerol (CBG) UDL

Cannabigerolic acid (CBGA) 4.76 ± 0.01

Total equivalent cannabigerol (TCBG) 4.45 ± 0.009 zUnder detection limit of 0.5 mg·g-1 of inflorescence dry weight. 384 yData are means ± SE (n = 22). 385

386



15

387

Table 2. The relationships between average photosynthetic photon flux density (APPFD) applied 388

during the flowering stage (81 days) and terpene potency in apical inflorescences of myrcene, 389

limonene and total terpenes, and the mean potency for terpenes with no APPFD treatment effects, of 390

Cannabis sativa L. ‘Stillwater’. 391

392

Terpene

Terpene potency

(mg·g-1 of inflorescence dry weight)

Mean z Regression equation y R2

Total terpenes Y = 0.00230 X + 8.57 0.320

Myrcene Y = 0.00142 X + 2.34 0.464

Limonene Y = 0.000326 X + 1.01 0.246

Alpha pinene 0.16z ± 0.01

Beta pinene 0.22 ± 0.01

Terpinolene UDLx

Linalool 0.53 ± 0.01

Terpineol 0.32 ± 0.02

Caryophyllene 2.9 ± 0.2

Humulene 0.65 ± 0.04

3-carene UDL

Cis-ocimene UDL

Eucalyptol UDL

Trans-ocimene UDL

Fenchol 0.22 ± 0.01

Borneol 0.03 ± 0.01

Valencene UDL

Cis-nerolidol UDL

Trans-nerolidol UDL

Guaiol UDL

Alpha-bisabolol 0.38 ± 0.03

Sabinene UDL zWhen there were no APPFD treatment effects on terpene potency, the means ± SE (n = 22) are 393

presented. 394 yLinear regression models for the APPFD treatment effects on terpene potency when P ≤ 0.05. 395 xUnder detection limit of 0.5 mg·g-1 of inflorescence dry weight. 396

397

3.3 Yield and Quality 398

Cannabis yield increased linearly from 116 to 519 g·m-2 (i.e., 4.5 times higher) as APPFD increased 399

from 120 to 1800 μmol·m-2·s-1 (Figure 6A). Note that yields in the present study are true oven-DWs. 400

Since fresh cannabis inflorescences are typically dried to 10 to 15% moisture content to achieve 401



16

optimum marketable quality (Leggett, 2006), yields in the present study can be easily adjusted 402

upwards to be comparable any desirable moisture level (e.g., by multiplying by 1.15 for 15% 403

moisture content). The harvest index increased linearly from 0.560 to 0.733 and (i.e., 1.3 times 404

higher) as APPFD increased from 120 to 1800 μmol·m-2·s-1 (Figure 6B). The AID increased linearly 405

from 0.0893 to 0.115 g·cm-3 (i.e., 1.3 times higher) as APPFD increased from 120 to 1800 μmol·m-406 2·s-1 (Figure 6C). 407

Cannabidiolic acid (CBDA) was the dominant cannabinoid in the dried inflorescences; however, 408

there were no APPFD treatment effects on the potency of any of the measured cannabinoids (Table 409

1). Due to linear increases in inflorescence yield with increasing LI, cannabinoid yield (g·m-2) 410

increased by 4.5 times as APPFD increased from 120 to 1800 μmol·m-2·s-1. Myrcene, limonene, and 411

caryophyllene were the dominant terpenes in the harvested inflorescences (Table 2). The potency of 412

total terpenes, myrcene, and limonene increased linearly from 8.85 to 12.7, 2.51 to 4.90, and 1.05 to 413

1.60 mg·g-1 inflorescence DW (i.e., 1.4, 2.0 and 1.5 times higher), respectively, as APPFD increased 414

from 120 to 1800 μmol·m-2·s-1. There were no APPFD effects on the potency of the other individual 415

terpenes. 416

4 DISCUSSION 417

4.1 Cannabis Inflorescence Yield is Proportional to Light Intensity 418

It was predicted that cannabis yield would exhibit a saturating response to increasing LI, thereby 419

signifying an optimum LI range for indoor cannabis production. However, the yield results of this 420

trial demonstrated cannabis’ immense plasticity for exploiting the incident lighting environment by 421

efficiently increasing marketable biomass up to extremely high – for indoor production – LIs (Figure 422

6A). Even under ambient CO2, the linear increases in yield indicated that the availability of PAR 423

photons was still limiting whole-canopy photosynthesis at APPFD levels as high as ≈ 1800 μmol·m-424 2·s-1 (i.e., DLI ≈ 78 mol·m-2·d-1). These results were generally consistent with the trends of other 425

studies reporting linear cannabis yield responses to LI (Eaves et al., 2020; Potter and Duncombe, 426

2012; Vanhove et al., 2011), although there is considerable variability in both relative and absolute 427

yield responses to LI in these prior works. The present study covered a broader range of LI, and with 428

much higher granularity, compared with other similar studies. 429

The lack of a saturating yield response at such high LI is an important distinction between cannabis 430

and other crops grown in controlled environments (Beaman et al., 2009; Fernandes et al., 2013; Oh et 431

al., 2009; Faust, 2003). This also means that the selection of an “optimum” LI for indoor cannabis 432

production can be made somewhat independently from its yield response to LI. Effectively, within 433

the range of practical indoor PPFD levels - the more light that is provided, the proportionally higher 434

the increase in yield will be. Therefore, the question of the optimum LI may be reduced to more 435

practical functions of economics and infrastructure limitations: basically, how much lighting capacity 436

can a grower afford to install and run? This becomes a trade-off between fixed costs which are 437

relatively unaffected by yield and profit (e.g., building lease/ownership costs including property tax, 438

licensing, and administration) and variable costs such as crop inputs (e.g., fertilizer, electricity for 439

lighting) and labor. Variable costs will obviously increase with higher LI but the fixed costs, on a per 440

unit DW basis, should decrease concomitantly with increasing yield (Vanhove et al., 2014). Every 441

production facility will have a unique optimum balance between facility costs and yield; but the yield 442

results in the present study can help cannabis cultivators ascertain the most suitable LI target for their 443

individual circumstances. Readers should be mindful that this study reports yield parameters as true 444



17

dry weights; marketable yield can be easily determined by factoring back in the desirable moisture 445

content of the inflorescence. For example, for a 400 g·m-2 of dry yield, the corresponding marketable 446

yield would be 440 g·m-2 at 10% moisture content (i.e., 400 x 1.10). 447

It is also important to appreciate that PPFD, which represents an instantaneous LI level, does not 448

provide a complete accounting of the total photon flux incident on the crop canopy throughout the 449

entire production cycle. While this LI metric is ubiquitous in the horticulture industry and may be 450

most broadly relatable to prior works, there is value in relating yield to the total photon flux received 451

by the crop. Historically, this has been done by relating yield to installed wattage on per area bases, 452

resulting in g·W-1 metric (Potter and Duncombe, 2012), which can be more fittingly converted to 453

yield per unit electrical energy input (g·kWh-1) by factoring in the photoperiod and length of the 454

production cycle (EMCDDA, 2013). However, since photosynthesis is considered a quantum 455

phenomenon, crop yield may be more appropriately related to incident (easily measured) or absorbed 456

photons and integrated over the entire production cycle (i.e., TLI, mol·m-2), in a yield metric that is 457

analogous to QY: g·mol-1. Versus using installed wattage, this metric has the advantage of negating 458

the effects of different fixture efficacy (μmol·J-1), which continues its upward trajectory, especially 459

with LEDs (Kusuma et al., 2020; Nelson and Bugbee, 2014). The present study did not directly 460

measure lighting-related energy consumption; however, installed energy flux (kWh·m-2) can be 461

estimated from TLI using the Lumigrow fixture’s efficacy rating: 1.29 and 1.80 μmol·J-1, from 462

Nelson and Bugbee (2014) and Radetsky (2018), respectively. Using the average of these values 463

(1.55 μmol·J-1), the conversion from TLI to energy flux becomes: mol·m-2 × 5.6 = kWh·m-2. At an 464

APPFD of 900 μmol·m-2·s-1 (i.e., TLI of 3149 mol·m-2), the model in Figure 6A predicts a yield of 465

303 g·m-2 which corresponds to an energy use efficacy of 0.54 g·kWh-1. For comparison, doubling 466

the LI to the highest APPFD used in this trial increases the yield by 70% but results in a ≈ 15% 467

reduction in energy use efficacy. It is up to each grower to determine the optimum balance between 468

variable (e.g., lighting infrastructure and energy costs) and fixed (e.g., production space) costs in 469

selecting a canopy level LI that will maximize profits. 470

4.2 Increasing Light Intensity Enhances Inflorescence Quality 471

Beyond simple yield, increasing LI also raised the harvest quality through higher apical inflorescence 472

(also called “chola” in the cannabis industry) density – an important parameter for the whole-bud 473

market – and increased ratios of inflorescence to total aboveground biomass (Figure 6B and 6C). 474

The linear increases in HI and AID with increasing LI both indicate shifts in biomass partitioning 475

more in favor of generative tissues; a common response in herbaceous plants (Poorter et al., 2019) 476

including cannabis (Hawley et al., 2018; Potter and Duncombe, 2012). The increases in these 477

attributes under high LI may also indirectly facilitate harvesting, as there is correspondingly less 478

unmarketable biomass to be processed and discarded, which is an especially labour-intensive aspect 479

of cannabis harvesting. 480

The terpene potency – comprised mainly of myrcene, limonene, and caryophyllene – increased by ≈ 481

25%, as APPFD increased from 130 to 1800 μmol·m-2·s-1 (Table 2), which could lead to enhanced 482

aromas and higher quality extracts (McPartland and Russo, 2001; Nuutinen, 2018). Conversely, total 483

cannabinoid yield increased in proportion with increasing inflorescence yield since there were no LI 484

treatment effects on cannabinoid potency (Table 1). Similarly, Potter and Duncombe (2012) and 485

Vanhove et al. (2011) found no LI treatment effects on cannabinoid potency (primarily THC in those 486

studies) and attributed increasing cannabinoid yield to enhanced biomass apportioning towards 487

generative tissues at higher LI. Other studies had contradictory results on the effects of LI on 488



18

potency. Hawley et al. (2018) did not find canopy position effects on THC or CBD potency in a 489

subcanopy lighting (SCL) trial, but they did find slightly higher cannabigerol potency in the upper 490

canopy in the control (high pressure sodium top-lighting only) and the Red-Green-Blue SCL 491

treatment, but not in the Red-Blue SCL treatment. While it is not possible to unlink spectrum from LI 492

in their results, the magnitude of the reported potency differences, both between canopy positions and 493

between lighting treatments, were relatively minor. Conversely, Namdar et al. (2018) reported what 494

appeared to be a vertical stratification on cannabis secondary metabolites, with highest potencies 495

generally found in the most distal inflorescences (i.e., closest to the light source, PPFD ≈ 600 496

μmol·m-2·s-1). They attributed this stratification to the localized LI at different branch positions, 497

which were reportedly reduced by ≥ 60% at lower branches vs. at the plant apex. However, given the 498

lack of LI treatment effects (over a much broader range of PPFDs) on cannabinoid potency in the 499

present study, it is likely that other factors were acting on higher-order inflorescences, such as 500

delayed maturation and reduced biomass allocation, that reduced potency in these tissues (Diggle, 501

1995; Hemphill et al., 1980). 502

4.3 Plasticity of Cannabis Leaf Morphology and Physiology Responses to LI and Over Time 503

The objectives of the photosynthesis and leaf morphology investigations in this study were twofold: 504

1) to address the knowledge gap in the relationships between localized cannabis leaf photosynthesis 505

and yield and 2) observe and report changes in physiology as the plant progresses through the 506

flowering ontogeny. 507

General morphological, physiological, and yield responses of plants are well documented across LI 508

gradients ranging from below compensation point to DLIs beyond 60 mol·m-2·d-1. Recently, the LI 509

responses of myriad plant attributes were compiled across a tremendous range species, ecotypes and 510

growing environments, and concisely reported them in the excellent review paper by Poorter et al. 511

(2019). The trends in their LI models align well with primary attributes measured in the present 512

study, including morphological parameters such as plant height and internode length (data not 513

shown), SLW (discussed below), and physiological parameters such as Fv/Fm, LNCER (i.e., 514

photosynthesis at growth light; Phot/AGL), and Asat (i.e., photosynthesis at saturating light; Phot/ASL). 515

In general, cannabis photosynthesis and yield responses to localized LI were linear across the APPFD 516

range of 120 to 1800 μmol·m-2·s-1. While these results are in agreement with the contemporary 517

literature on cannabis (Bauerle et al., 2020; Chandra et al., 2008; 2015; Eaves et al., 2020; Potter and 518

Duncombe, 2012), we also showed substantial chronological dependencies on leaf photosynthetic 519

indices. 520

By surveying the photosynthetic parameters of the upper cannabis canopy across a broad range of 521

LPPFDs and over multiple timepoints during the generative phase, we saw evidence of both 522

acclimation and early senescence as the crop progressed through its ontogeny. At the beginning of 523

the trial, the plants were abruptly transitioned from a uniform PPFD (425 μmol·m-2·s-1) and 18-h 524

photoperiod (i.e., 27.5 mol·m-2·d-1) and subjected to a much shorter photoperiod (12-h) and an 525

enormous range of LI (120 to 1800 μmol·m-2·s-1), resulting in DLIs ranging from 5.2 to 78 mol·m-526 2·d-1. Further, on a DLI-basis, approximately 1/3 of the plants were exposed to lower LIs in the 527

flowering vs. vegetative phase (i.e., APPFD < 640 μmol·m-2·s-1). These sudden transitions in both LI 528

and photoperiod resulted in substantive changes in the plants’ lighting environment at the start of the 529

trial, stimulating various morphological and physiological adaptations with differing degrees of 530

plasticity. The leaves measured in week 1 developed and expanded during the prior vegetative phase 531

under a different lighting regimen (LI and photoperiod). The leaves measured in week 5 were 532



19

developed under their respective LPPFDs during a period characterized by slowing vegetative growth 533

and transitioning to flower development. The leaves measured in week 9 would have also developed 534

under their respective LPPFDs, but since cannabis vegetative growth greatly diminishes after the first 535

five weeks in 12-h days (Potter, 2014), these tissues were physiologically much older than the leaves 536

measured in week 5, with concomitant reductions in photosynthetic capacity (Bauerle et al., 2020; 537

Bielczynski et al., 2017). 538

These differences in leaf physiological age, plant ontogeny, and localized lighting environments 539

during leaf expansion vs. measurement resulted in notable temporal variability in leaf-level LI 540

responses. In week 1, there were no LI treatment effects on Asat and the slopes of the LSP, LNCER, 541

and Fv/Fm were shallower in weeks 5 and 9. The comparatively lower LI responses in week 1 were 542

likely due to the reduced adaptive plasticity that mature foliar tissues have vs. leaves that developed 543

under a new lighting regime (Sims and Pearcy, 1992). Further, Y-intercepts for the Asat, LSP, and 544

LNCER models were higher in week 1 than weeks 5 and 9, which may be partly due to the higher LI 545

(amplified by the longer photoperiod) that the leaves developed under, during the latter part of the 546

vegetative phase. Further, the Asat, LSP, and LNCER models in weeks 5 and 9 have comparable 547

slopes, but there is a vertical translation in the respective models, resulting week 9 models having 548

substantially lower Y-intercepts (i.e., approximately half) for these parameters. The interplay of 549

physiological age of foliage and plant ontogeny (i.e., onset of senescence) on the diminished 550

photosynthetic capacity of the leaves in week 9 is unknown, but the dynamic temporal nature of 551

cannabis photosynthesis (during flowering) is manifest in these models. 552

Given these impacts of physiological age and light history, we posit that cannabis leaf photosynthesis 553

cannot be used as a stand-alone gauge for predicting yield. Chandra et al. (2008) and Chandra et al. 554

(2015) provided insight into the substantial capacity for drug-type strains of indoor grown cannabis 555

leaves to respond to LI; and the results of these trials are much lauded in the industry as evidence that 556

maximum photosynthesis and yields will be reached under canopy-level PPFDs of ≈1500 μmol·m-557 2·s-1. However, the 400 to 500 μmol·m-2·s-1 increments in LPPFD does not provide sufficient 558

granularity (particularly at low LI) to reliably model the LRCs, thus no models were provided. 559

Further, the LRCs were made on leaves of varying and unreported physiological ages, from plants 560

exposed to a vegetative photoperiod (18-h), and acclimated to unspecified localized LI (a canopy-561

level PPFD of 700 μmol·m-2·s-1 was indicated in Chandra et al., 2015). The strong associations 562

between a tissue’s light history and its photosynthesis responses to LI, demonstrated in this trial and 563

by others (Björkman, 1981), represent a major shortcoming of using leaf LI response models to infer 564

crop growth and yield. To illustrate, Figure 2 shows LRCs of leaves from a single cultivar, at similar 565

physiological ages (week 5 after transition to 12-h photoperiod) but acclimated to disparate LPPFDs: 566

91 and 1238 μmol·m-2·s-1. The relative difference in LNCER at higher LIs (≈ 50%) between these 567

two curves is representative of the potential uncertainty due to just one of the uncontrolled 568

parameters (LNCER) in these prior works. Differing physiological ages of tissues at the time of 569

measurement may have conferred an even larger degree of uncertainty in the magnitude of leaf 570

responses to LI (Bauerle et al., 2020) than leaf light history. Consideration must also be given to the 571

different life stages of a photoperiodic crop (i.e., vegetative vs. generative) and the inherent impact 572

that day length imbues on the total daily PAR exposure (i.e., DLI) which can correlate better to crop 573

yield than PPFD. Further, for a given DLI, yields are higher under longer photoperiod (Vlahos et al., 574

1991; Zhang et al., 2018), ostensibly due to their relative proximity to their maximum QY (Ohyama 575

et al., 2005). A final distinction between leaf photosynthesis and whole plant yield responses to LI is 576

the saturating LI: the LSP for leaf photosynthesis were substantially lower than the LSP for yield, 577

which remains undefined due to the linearity of the light response model. 578



20

Newly-expanded leaves, especially in herbaceous species, are able to vary their leaf size, thickness 579

and chlorophyll content in response to LPPFD in order to balance myriad factors such as internal and 580

leaf surface gas exchange (CO2 and H2O), internal architecture of the light-harvesting complexes, and 581

resistance to photoinhibition (Björkman, 1981). In the present study, the effects of LI on leaf 582

morphology was only evaluated in week 5, when the crop was still actively growing vegetative 583

biomass. Reductions in SLW (i.e., increases in specific leaf area, SLA) in response to increasing LI 584

are abundant in the literature (Fernandes et al., 2013; Gratani, 2014; Sims and Pearcy, 1992). In 585

particular, Poorter et al. (2019) reported a saturating response of SLW [also known as leaf mass (per) 586

area; LMA] to LI across 520 species (36% of which were herbaceous plants), however much of their 587

data was at DLIs lower than the minimum DLI in the present study (5.2 mol·m-2·d-1), which affected 588

the shape of their SLW response model to LI. Across similar DLI ranges, the average increase in 589

SLW across 520 species was 1.7 X in Poorter et al., (2019) vs 1.6 X in the present study, indicating 590

that cannabis SLW responses to LI are consistent with normal trends for this parameter. 591

The lack of LI treatment effects on CCI are also consistent with other studies that have shown that 592

area-based chlorophyll content is fairly stable across a broad range of LIs (Poorter et al., 2019; 593

Björkman, 1981), despite substantial variability in photosynthetic efficiency. However, since there 594

were LI treatment effects on SLW, chlorophyll content on leaf volume or mass bases would likely 595

have reduced under higher LI. The positional effects on CCI (i.e., higher in upper vs. lower canopy) 596

were probably due to the interplay between self-shading and advancing physiological age of the 597

lower leaves (Bauerle et al., 2020). The temporal effects on CCI, which was higher in week 1 vs. 598

weeks 5 and 9, in both upper and lower leaves, may have been due to changes in QY over the life-599

cycle of the crop. Bugbee and Monje (1992) presented a similar trend high QY during the active 600

growth phase of a 60-d crop cycle of wheat, followed by a reduction in QY at the onset of senescence 601

(i.e., shortly before harvest). The decline in chlorophyll content in the latter phase of the production 602

cycle probably contributed to the reductions in the photosynthetic parameters (e.g., Asat, LSP, 603

LNCER) of the tissues measured in week 9 vs. week 5. 604

Overall, the impact that increasing LI had on cannabis morphology and yield were captured 605

holistically in the plant sketches in Figure 5, which shows plants grown under higher LIs had shorter 606

internodes, smaller leaves, and much larger and denser inflorescences (resulting in higher HI), 607

especially at the plant apex. Like many other plant species, we have found that cannabis has immense 608

plasticity to rapidly acclimate its morphology and physiology, both at leaf- and whole plant-levels, to 609

changes in the growing lighting environment. Therefore, in order reliably predict cannabis growth 610

and yield to LI, it is necessary to grow plants under a broad range of LIs through their full ontological 611

development, as was done in this study. Without knowing the respective tissues’ age and light 612

history, instantaneous light response curves at leaf-, branch-, or even canopy-levels cannot reliably 613

predict yield. 614

5 CONCLUSIONS 615

We have shown an immense plasticity for cannabis to respond to increasing LI; in terms of 616

morphology, physiology (over time), and yield. The temporal dynamics in cannabis leaf acclimations 617

to LI have also been explored, addressing some knowledge-gaps in relating cannabis photosynthesis 618

to yield. The results also indicate that the relationship between LI and cannabis yield does not 619

saturate within the practical limits of LI used in indoor production. Increasing LI also increased HI 620

and the size and density of the apical inflorescence; both markers for increasing quality. However, 621

there were no and minor LI treatment effects on potency of cannabinoids and terpenes, respectively. 622



21

This means that growers may be able to vastly increase yields by increasing LI but maintain a 623

relatively consistent secondary metabolite profile in their marketable products. Ultimately, the 624

selection of the economic optimum canopy-level LI for a given commercial production system 625

depends on many interrelated factors. 626

Future research should expand to multiple cultivars of both indica- and sativa-dominant biotypes. 627

Further, since plant yield responses to elevated CO2 can mirror the responses to elevated LI, the 628

combined effects of CO2 and LI should be investigated on cannabis yield with an in-depth cost-629

benefit analysis of the optimum combination of these two input parameters. 630

6 ABBREVIATIONS 631

NCER; Net CO2 exchange rate, PPFD; photosynthetic photon flux, Asat; light-saturated NCER, LSP; 632

light saturation point, QY; maximum quantum yield, CCI; chlorophyll content index, SLW; specific 633

leaf weight, LED; light emitting diode, DLI; daily light integral, PAR; photosynthetically active 634

radiation, DW; dry weight, SD; standard deviation, SE; standard error, RH; relative humidity, Δ9-635

THC; Δ-9-tetrahydrocannabinol, Δ9-THCA; Δ-9-tetrahydrocannabinolic acid, TΔ9-THC; total 636

equivalent Δ9-tetrahydrocannabinol, CBD; cannabidiol, TCBD; total equivalent cannabidiol, CBG; 637

cannabigerol, CBGA; cannabigerolic acid, TCBG; total equivalent cannabigerol. 638

6.1 Non-Standard Abbreviations 639

LPPFD; localized PPFD at the measured leaf, APPFD; average PPFD at the plant apex integrated over 640

time, LNCER; NCER at LPPFD, AID; apical inflorescence density, LI; light intensity, HI; harvest 641

index, TLI; total light integral, LRC; light response curve, CB; deep-water culture basin, UDL; under 642

detection limit 643

7 AUTHOR CONTRIBUTIONS 644

All authors contributed to the experimental design. VRM and DL performed the experiment, 645

collected and analyzed the data. DL, VRM and YZ wrote and revised the manuscript. All authors 646

approved the final manuscript. 647

8 FUNDING 648

This work was funded by Natural Sciences and Engineering Research Council of Canada. Green 649

Relief provided the research facility, cannabis plants, experimental materials, and logistical support. 650

9 ACKNOWLEDGEMENTS 651

We thank Dr. Nigel Gale for contributing to the experimental design. We thank Derek Bravo, Tim 652

Moffat, and Dane Cronin for technical support throughout the experiment. 653

10 CONTRIBUTION TO THE FIELD STATEMENT 654

Recent legalization of cannabis in many regions world-wide has provoked demand for scientific 655

research to improve cannabis production. Several studies have established models to estimate 656

cannabis floral yield response to light intensity. While these works have shown cannabis’ immense 657

capacity for converting light into marketable biomass, their models cannot be directly utilized by 658



22

cannabis producers without copious assumptions. Therefore, we evaluate the impact of a refined 659

range of light intensities (testing the lower and upper limits of practical light intensities used in 660

cannabis production) on cannabis physiology, morphology, yield, and quality. We demonstrate the 661

extraordinary plasticity of cannabis’ physiological, morphological and yield responses to increasing 662

light intensity. We also demonstrate that leaf-level photosynthetic responses to light intensity vary 663

substantially with leaf age and light history. Therefore, leaf- and plant-level photosynthetic responses 664

cannot reliably predict cannabis yield responses to light intensity. This research will assist growers 665

in making informed decisions about the optimum light intensity to use for their production systems. 666

667

11 REFERENCES 668

Backer, R., Schwinghamer, T., Rosenbaum, P., McCarty, V., Eichhorn Bilodeau, S., Lyu, D., 669

Ahmed, M.B., Robinson, G., Lefsrud, M., Wilkins, O., Smith, D.L. (2019). Closing the yield 670

gap for cannabis: A meta-analysis of factors determining cannabis yield. Front. Plant Sci. 671

10:495. doi: 10.3389/fpls.2019.00495. 672

673

Bauerle, W.L., McCullough, C., Iversen, M., Hazlett, M. (2020). Leaf age and position effects on 674

quantum yield and photosynthetic capacity in hemp crowns. Plants 9:271. doi: 675

10.3390/plants9020271. 676

677

Beaman, A.R., Gladon, R.J., Schrader, J.A. (2009). Sweet basil requires an irradiance of 500 678

μmol·m-2·s-1 for greatest edible biomass production. HortScience 44:64–67. doi: 679

10.21273/HORTSCI.44.1.64. 680

681

Bielczynski, L.W., Łaçki, M.K., Hoefnagels, I., Gambin, A., Croce, R. (2017). Leaf and plant age 682

affects photosynthetic performance and photoprotective capacity. Plant Physiol. 175:1634–683

1648. doi: 10.1104/pp.17.00904. 684

685

Björkman, O. (1981). "Responses to different quantum flux densities," in: Encyclopedia of Plant 686

Physiology: New Series: Vol. 12A, Physiological Plant Ecology 1, eds. L. Lange, P.S. Nobel, 687

C.B. Osmond, and H. Ziegler (Heidelberg, Berlin: Springer-Verlag), 57-107. 688

689

Bredmose, N. (1993). Effects of year-round supplementary lighting on shoot development, flowering 690

and quality of two glasshouse rose cultivars. Sci. Hortic. 54:69–85. doi: 10.1016/0304-691

4238(93)90084-4. 692

693

Bredmose, N. (1994). Biological efficiency of supplementary lighting on cut roses the year round. 694

Sci. Hortic. 59:75–82. doi: 10.1016/0304-4238(94)90094-9. 695

696

Bugbee, B., Monje, O. (1992). The limits of crop productivity. Bioscience. 42:494–502. doi: 697

10.2307/1311879. 698

699

Carvalho, F.E.L., Ware, M.A., Ruban, A. V. (2015). Quantifying the dynamics of light tolerance in 700

Arabidopsis plants during ontogenesis. Plant Cell Environ. 38:2603–2617. doi: 701

10.1111/pce.12574. 702

703



23

Chandra, S., Lata, H., Khan, I.A., Elsohly, M.A. (2008). Photosynthetic response of Cannabis sativa 704

L. to variations in photosynthetic photon flux densities, temperature and CO2 conditions. 705

Physiol. Mol. Biol. Plants 14:299–306. doi: 10.1007/s12298-008-0027-x. 706

707

Chandra, S., Lata, H., Khan, I.A., ElSohly, M.A. (2011). Photosynthetic response of Cannabis sativa 708

L., an important medicinal plant, to elevated levels of CO2. Physiol. Mol. Biol. Plants 709

17:291–295. doi: 10.1007/s12298-011-0066-6. 710

711

Chandra, S., Lata, H., Mehmedic, Z., Khan, I.A., ElSohly, M.A. (2015). Light dependence of 712

photosynthesis and water vapor exchange characteristics in different high Δ9-THC yielding 713

varieties of Cannabis sativa L. J. Appl. Res. Med. Aromat. Plants 2:39–47. doi: 714

10.1016/J.JARMAP.2015.03.002. 715

716

Delgado, E., Parry, M.A.J., Lawlor, D.W., Keys, A.J., Medrano, H. (1993). Photosynthesis, ribulose-717

1,5-bisphosphate carboxylase and leaf characteristics of Nicotiana tabacum L. genotypes 718

selected by survival at low CO2 concentrations. J. Exp. Bot. 44:1–7. doi: 10.1093/jxb/44.1.1. 719

720

Diggle, P.K. (1995). Architectural effects and the interpretation of patterns of fruit and seed 721

development. Annu. Rev. Ecol. Syst. 26:531–552. doi: 722

10.1146/annurev.es.26.110195.002531. 723

724

Eaves, J., Eaves, S., Morphy, C., Murray, C. (2020). The relationship between light intensity, 725

cannabis yields, and profitability. Agron. J. 112:1466–1470. doi: 10.2139/ssrn.3310456. 726

727

European Monitoring Centre for Drugs and Drug Addiction. (2013). Cannabis Production and 728

Markets In Europe. Lisbon, Portugal: EMCDDA Insights, No. 12. 729

730

Evergreen Economics. (2016). Cannabis Agriculture Energy Demand Study Final Report. San 731

Diego, CA: San Diego Gas & Electric Company. 732

733

Faust, J.E. (2003). "Light," in Ball Redbook, ed. D. Hamrick (Batabia, IL: Ball Publishing), 71–84. 734

735

Fernandes, V.F., de Almeida, L.B., da Feijó, E.V.R.S., da Silva, D.C., de Oliveira, R.A., Mielke, 736

M.S., do Costa, L.C.B. (2013). Light intensity on growth, leaf micromorphology and essential 737

oil production of Ocimum gratissimum. Brazilian J. Pharmacogn. 23:419–424. doi: 738

10.1590/S0102-695X2013005000041. 739

740

Gratani, L. (2014). Plant phenotypic plasticity in response to environmental factors. Adv. Bot. 741

2014:1–17. doi: 10.1155/2014/208747. 742

743

Hawley, D., Graham, T., Stasiak, M., Dixon, M. (2018). Improving cannabis bud quality and yield 744

with subcanopy lighting. HortScience 53:1593–1599. doi: 10.21273/hortsci13173-18. 745

746

Hemphill, J.K., Turner, J.C., Mahlberg, P.G. (1980). Cannabinoid content of individual plant organs 747

from different geographical strains of Cannabis sativa L. J. Nat. Prod. 43:112–122. doi: 748

10.1021/np50007a009. 749

750

Jones-Baumgardt, C., Llewellyn, D., Ying, Q., Zheng, Y. (2019). Intensity of sole-source light-751



24

emitting diodes affects growth, yield, and quality of Brassicaceae microgreens. HortScience. 752

54:1168–1174. doi: 10.21273/HORTSCI13788-18. 753

754

Kirschbaum, M.U.F. (2011). Does enhanced photosynthesis enhance growth? Lessons learned from755

CO2 enrichment studies. Plant Physiol. 155:117–124. doi: 10.1104/pp.110.166819. 756

757

Kreyling, J., Schweiger, A.H., Bahn, M., Ineson, P., Migliavacca, M., Morel-Journel, T., 758

Christiansen, J.R., Schtickzelle, N., Larsen, K.S. (2018). To replicate, or not to replicate - that 759

is the question: How to tackle nonlinear responses in ecological experiments. Ecol. Lett. 760

21:1629–1638. doi: 10.1111/ele.13134. 761

762

Kusuma, P., Pattison, P.M., Bugbee, B. (2020). From physics to fixtures to food: Current and 763

potential LED efficacy. Hortic. Res. 7:56. doi: 10.1038/s41438-020-0283-7. 764

765

Leggett, T. (2006). A review of the world cannabis situation. Bull. Narc. 58: 1–155. 766

767

Lobo, F. A., de Barros, M.P., Dalmagro, H.J., Dalmolin, Â.C., Pereira, W.E., de Souza, É.C., 768

Vourlitis, G.L., Rodríguez Ortíz, C.E. (2013). Fitting net photosynthetic light-response curves 769

with Microsoft Excel - a critical look at the models. Photosynthetica. 51:445–456. doi: 770

10.1007/s11099-013-0045-y. 771

772

Lydon, J., Teramura, A.H., Coffman, C.B. (1987). UV-B radiation effects on photosynthesis, growth 773

and cannabinoid production of two Cannabis sativa chemotypes. Photochem. Photobiol. 774

46:201–206. doi: 10.1111/j.1751-1097.1987.tb04757.x. 775

776

Marcelis, L.F.M., Broekhuijsen, A.G.M., Meinen, E., Nijs, E.M.F.M., Raaphorst, M.G.M. (2006). 777

Quantification of the growth response to light quantity of greenhouse grown crops. Acta 778

Hortic. 711:97–103. doi: 10.17660/actahortic.2006.711.9. 779

780

McPartland, J.M. (2018). Cannabis systematics at the levels of family, genus, and species. Cannabis 781

Cannabinoid Res. 3:203–212. doi: 10.1089/can.2018.0039. 782

783

Mills, E. (2012). The carbon footprint of indoor Cannabis production. Energy Policy. 46: 58–67. doi: 784

10.1016/J.ENPOL.2012.03.023. 785

786

Murchie, E., Hubbart, S., Chen, Y., Peng, S., Horton, P. (2002). Acclimation of rice photosynthesis787

to irradiance under field conditions. Plant Physiol. 130:1999–2010. doi: 10.1104/pp.011098. 788

789

Murchie, E., Lawson, T. (2013). Chlorophyll fluorescence analysis: A guide to good practice and 790

understanding some new applications. J. Exp. Bot. 64:3983–3998. doi: 10.1093/jxb/ert208. 791

792

Namdar, D., Mazuz, M., Ion, A., Koltai, H. (2018). Variation in the compositions of cannabinoid and 793

terpenoids in Cannabis sativa derived from inflorescence position along the stem and 794

extraction methods. Ind. Crops Prod. 113:376–382. doi: 10.1016/j.indcrop.2018.01.060. 795

796

Nelson, J.A., Bugbee, B. (2014). Economic analysis of greenhouse lighting: Light emitting diodes vs. 797

high intensity discharge fixtures. PLoS One. 9(6). doi: 10.1371/journal.pone.0099010. 798

799



25

Niinemets, Ü., Keenan, T.F. (2012). Measures of light in studies on light-driven plant plasticity in 800

artificial environments. Front. Plant Sci. 3:156. doi: 10.3389/fpls.2012.00156. 801

802

Nuutinen, T. (2018). Medicinal properties of terpenes found in Cannabis sativa and Humulus 803

lupulus. Eur. J. Med. Chem. 157:198–228. doi: 10.1016/j.ejmech.2018.07.076. 804

805

Oh, W., Cheon, H., Kim, K.S., Runkle, E.S. (2009). Photosynthetic daily light integral influences 806

flowering time and crop characteristics of Cyclamen persicum. HortScience. 44:341–344. doi: 807

10.21273/HORTSCI.44.2.341. 808

809

Ohyama, K., Manabe, K., Omura, Y., Kozai, T., Kubota, C. (2005). Potential use of a 24-hour 810

photoperiod (continuous light) with alternating air temperature for production of tomato plug 811

transplants in a closed system. HortScience. 40:374–377. doi: 10.21273/HORTSCI.40.2.374. 812

813

Poorter, H., Niinemets, Ü., Ntagkas, N., Siebenkäs, A., Mäenpää, M., Matsubara, S., Pons, T. (2019). 814

A meta‐analysis of plant responses to light intensity for 70 traits ranging from molecules to 815

whole plant performance. New Phytol. 223:1073–1105. doi: 10.1111/nph.15754. 816

817

Posada, J.M., Sievänen, R., Messier, C., Perttunen, J., Nikinmaa, E., Lechowicz, M.J. (2012). 818

Contributions of leaf photosynthetic capacity, leaf angle and self-shading to the maximization 819

of net photosynthesis in Acer saccharum: a modelling assessment. Ann. Bot. 110:731–741. 820

doi: 10.1093/aob/mcs106. 821

822

Potter, D.J. (2014). A review of the cultivation and processing of cannabis (Cannabis sativa L.) for 823

production of prescription medicines in the UK. Drug Test. Anal. 6:31–38. doi: 824

10.1002/dta.1531. 825

826

Potter, D.J., Duncombe, P. (2012). The effect of electrical lighting power and irradiance on indoor-827

grown cannabis potency and yield. J. Forensic Sci. 57:618–622. doi: 10.1111/j.1556-828

4029.2011.02024.x. 829

830

Radetsky, L.C. (2018). LED and HID Horticultural Luminaire Testing Report Prepared For Lighting 831

Energy Alliance Members and Natural Resources Canada. Troy, NY: Rensselaer Polytechnic 832

Institute. 833

834

Sen, A., Wyonch, R. (2018). Cannabis Countdown: Estimating the Size Of Illegal Markets and Lost 835

Tax Revenue Post-Legalization. Toronto, ON: C.D. Howe Institute. 836

837

Sims, D.A., Pearcy, R.W. (1992). Response of leaf anatomy and photosynthetic capacity in Alocasia 838

macrorrhiza (Araceae) to a transfer from low to high light. Am. J. Bot. 79:449-455. doi: 839

10.2307/2445158. 840

841

Singsaas, E.L., Ort, D.R., DeLucia, E.H. (2001). Variation in measured values of photosynthetic 842

quantum yield in ecophysiological studies. Oecologia 128:15–23. doi: 843

10.1007/s004420000624. 844

845

Terashima, I., Hikosaka, K. (1995). Comparative ecophysiology of leaf and canopy photosynthesis. 846

Plant. Cell Environ. 18:1111–1128. doi: 10.1111/j.1365-3040.1995.tb00623.x. 847



26

848

Toonen, M., Ribot, S., Thissen, J. (2006). Yield of illicit indoor cannabis cultivation in the 849

Netherlands. J. Forensic Sci. 51:1050–1054. doi: 10.1111/j.1556-4029.2006.00228.x. 850

851

United Nations Office on Drugs and Crime. (2019). World Drug Report 2019. United Nations: New 852

York, NY. 853

854

Vanhove, W., Van Damme, P., Meert, N. (2011). Factors determining yield and quality of illicit 855

indoor cannabis (Cannabis spp.) production. Forensic Sci. Int. 212:158–163. doi: 856

10.1016/J.FORSCIINT.2011.06.006. 857

858

Vanhove, W., Surmont, T., Van Damme, P., De Ruyver, B. (2014). Filling in the blanks. An 859

estimation of illicit cannabis growers’ profits in Belgium. Int. J. Drug Policy 25:436–443. doi: 860

10.1016/j.drugpo.2014.01.020. 861

862

Vlahos, J.C., Heuvelink, E., Martakis, G.F.P. (1991). A growth analysis study of three Achimenes 863

cultivars grown under three light regimes. Sci. Hortic. 46:275–282. doi: 10.1016/0304-864

4238(91)90050-9. 865

866

Zhang, X., He, D., Niu, G., Yan, Z., Song, J. (2018). Effects of environment lighting on the growth, 867

photosynthesis, and quality of hydroponic lettuce in a plant factory. Int. J. Agric. Biol. Eng. 868

11:33–40. doi: 10.25165/j.ijabe.20181102.3420. 869

870

Zheng, Y. (2020). Soilless production of drug-type Cannabis sativa. Acta Hortic. (In press.). 871

872

Zheng, Y., Blom, T., Dixon, M. (2006). Moving lamps increase leaf photosynthetic capacity but not 873

the growth of potted gerbera. Sci. Hortic. 107:380–385. doi: 10.1016/j.scienta.2005.09.004. 874

875

876



Documents

Cannabis yield, potency, and leaf photosynthesis respond